Backends on Vercel: Python, containers, and Daemons

Vercel has announced general availability for Python backend support, allowing developers to deploy Python frameworks such as FastAPI, Flask, and Django using standard CPython and PyPI dependencies. The platform handles framework detection, builds, and scaling, while maintaining the familiar frontend development experience including local development, previews, and production deployments. Over 150,000 teams, including engineers at OpenAI, Polymarket, and Exa, already run Python backends on Vercel.

Container support in Vercel Functions enables deployment of applications written in any language, packaged with required OS libraries and executables. Since entering Public Beta, 5,000 teams have completed 150,000 container deployments, which scale with traffic and use Active CPU pricing to avoid charges during idle periods such as API responses or database queries. Containers deploy as Vercel Functions on Fluid compute, integrating with Vercel Services for backend development alongside frontend code.

Vercel Daemons, now in Public Beta, allows deployment of containers as continuously running background workers on Fluid compute. Users can specify the number of instances per region, with Vercel maintaining capacity by restarting processes if they exit or crash. Daemons supports pull-based workloads, where code requests work from external systems and processes it as available, such as consuming queue messages or fetching pending jobs from a database.

The expanded backend capabilities include SDKs for Python to integrate with Vercel products like Blob for file storage, Queues for background jobs, Workflows for durable steps, Sandbox for untrusted code, and the AI SDK for streaming model responses. Delphi, an AI digital minds company, migrated its Python backend from AWS ECS to Vercel, enabling a team of ten engineers with no dedicated infrastructure role to ship to production over 100 times daily using Queues for background jobs.